File size: 4,115 Bytes
87bb323 352b288 87bb323 f84a6ca 1dcd225 352b288 7c2871c 7eda558 c7ccb6a 1dde14f 0d99b6e 9d10737 22b6335 2476bfa 0d9ad48 2806269 7c2871c 352b288 7e916d4 54b3a9a 7e916d4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
---
license: cc-by-sa-3.0
tags:
- art
---
<div align="center"><font size="8"><b>starlake</b><font></div>
<div align="center"><font size="5"><b>这是一个试图同时兼顾风景和人物的模型</b><font></div>
<font size="5">在无人物描述时,不会出现通常模型倾向于添加人物的情况
![xyz_grid-0002-4154837933.png](https://s3.amazonaws.com/moonup/production/uploads/1678859211392-640cae30536d9fe0f002ee62.png)
<font color="red">PS:负面可以直接使用emb:easynegative或者verybadimagenegative_v1.3,这两个都是影响强烈的emb模型,可以降权使用,比如(verybadimagenegative_v1.3:0.8),否则会给模型带来一定的过拟合光影,当不需要过度光影的情况下,这是个好方法。</font>
![xyz_grid-0033-1174306601.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/MMccIAkPYzm7njo4TMw4p.png)
![xyz_grid-0035-84983774.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/gi2RG1n9GL63zDBlwQUd0.png)
___________________________
已完成5.0版本,5.0软化了过拟合因素,不加入详细的提示词不会自主添加元素,同时也增强了读tag能力。
本次训练全程添加金字塔噪声,5.0可以直接出明暗强烈的图。
新的例图展示如下:
人物
![00010-1710471201.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/770CynRxico7AEXpRsMWl.png)
![00747-285416818.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/2yXSzdWgGJhTkf-Y-UfG2.png)
![00069-3762213293.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/Qeqm5JjuUpZ_u97Mvru_b.png)
场景
![00064-1030430370.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/v8WfCsaEutiCT4OOFfJUv.png)
![00072-1703637542.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/bUyxFMVEUq5jfMPkI38NL.png)
![00748-824635104.png](https://cdn-uploads.huggingface.co/production/uploads/640cae30536d9fe0f002ee62/DYBsyLy1g0AHrRFwvnGlc.png)
___________________________
3.5版本已上传。
![00126-2780910881.png](https://s3.amazonaws.com/moonup/production/uploads/640cae30536d9fe0f002ee62/TIbqPtSkH-bF_Eh53wYBY.png)
![00026-3840530172.png](https://s3.amazonaws.com/moonup/production/uploads/640cae30536d9fe0f002ee62/R3HhLHFoKStUb4M9RegQp.png)
![00048-1899924086.png](https://s3.amazonaws.com/moonup/production/uploads/640cae30536d9fe0f002ee62/g-7lIfJ3-D0Ly9SxrmhGt.png)
![00091-1273521536.png](https://s3.amazonaws.com/moonup/production/uploads/640cae30536d9fe0f002ee62/mxaR4pMPruWCUmZVi6TE5.png)
![00088-2351873475.png](https://s3.amazonaws.com/moonup/production/uploads/640cae30536d9fe0f002ee62/2OpHleA3vDFS80i6VEAq7.png)
以下为2.0旧测试
___________________________
远景测试:
![00255-3579052070.png](https://s3.amazonaws.com/moonup/production/uploads/1678716145346-640cae30536d9fe0f002ee62.png)
![grid-0005.png](https://s3.amazonaws.com/moonup/production/uploads/1678716104574-640cae30536d9fe0f002ee62.png)
![00254-2280245862.png](https://s3.amazonaws.com/moonup/production/uploads/1678716181398-640cae30536d9fe0f002ee62.png)
人物测试:
![grid-0016.png](https://s3.amazonaws.com/moonup/production/uploads/1678716259214-640cae30536d9fe0f002ee62.png)
![grid-0001.png](https://s3.amazonaws.com/moonup/production/uploads/1678716224999-640cae30536d9fe0f002ee62.png)
![grid-0009.png](https://s3.amazonaws.com/moonup/production/uploads/1678716241591-640cae30536d9fe0f002ee62.png)
![grid-0011.png](https://s3.amazonaws.com/moonup/production/uploads/1678716307863-640cae30536d9fe0f002ee62.png)
其他例图:
![00299-279408421.png](https://s3.amazonaws.com/moonup/production/uploads/1678716327013-640cae30536d9fe0f002ee62.png)
![00256-3983416131.png](https://s3.amazonaws.com/moonup/production/uploads/1678716354044-640cae30536d9fe0f002ee62.png)
<font size="5"><b>starlake为未融合版本,存在较多人体结构问题,可用于融合。</b><font>
<b>starlake-2.0为通过模型融合修正后的版本,可直接使用。</b> |